Pain in women with knee and/or hip osteoarthritis is related to adipose tissue dysfunction - data from the khoala cohort

Purpose: Beyond the link between metabolic diseases and osteoarthritis (OA) risk, some studies have suggested an association between metabolic syndrome or visceral obesity and OA-related pain. Here, we investigated whether biomarkers of adipose tissue dysfunction or of systemic inflammation could be associated with OA-related pain. Methods: We cross-sectionally analyzed patients with knee and/or hip OA at inclusion in the 863 patients included in the national multicentric KHOALA cohort (NCT00481338). We used visual analogic scale (VAS) for pain, the Western Ontario and McMaster Universities Arthritis Index (WOMAC) and Osteoarthritis Knee and Hip Quality of Life (OAKHQOL) pain subscores. For WOMAC and VAS pain, higher scores reflect more symptoms, while it is the opposite for OAKHQOL. At inclusion, we measured in the serum ultra-sensitive C reactive protein (usCRP), reflecting systemic inflammation, leptin and total adiponectin for calculation of leptin:adiponectin ratio (LAR), a marker of adipose tissue dysfunction associated with central adiposity, high-molecular-weight adiponectin, visfatin and apolipoproteins (apoA1, apoB100). Univariate and multivariate analyses using stepwise linear regression models were performed to search for correlation between pain assessments and these biomarkers, with adjustment on age and Kellgren-Lawrence score and with stratification by sex. Results: In 596 women with hip and/or knee OA, multivariate analyses indicated that higher pain intensity was associated with higher LAR (VAS pain: β=0.49; p=0.0001, OAKHQOL pain: β=-0.46; p=0.0002, WOMAC pain: β=0.3; p=0.002) in the whole group. Analyzing knee and hip OA populations separately, these correlations were found for both joints in multivariate analyses (knee OA: VAS pain β = 0.55 p = 0.0001; OAKHQOL pain: β = -0.50 p = 0.0002; WOMAC pain β = 0.30 p = 0.006; hip OA: VAS pain β = 0.64 p = 0.03; OAKHQOL pain: β = -0.55 p = 0.04). Pain intensity correlated also with usCRP level (VAS pain: β= 0.27; p=0.02, OAKHQOL pain: β =-0.31; p=0.01) and Kellgren-Lawrence score. Serum visfatin, high-molecular-weight adiponectin, Apo A1 and Apo B100 levels were not related to pain level, whatever the score used. In 267 men, no correlation between biomarkers and OA pain was found. Conclusions: Serum LAR and usCRP level are associated with pain level, independently of radiographic structural severity in women with hip and/or knee OA, emphasizing the role of adipose tissue dysfunction and of systemic inflammation in pain experience in the OA female population.